Science of the Total Environment 515-516 (2015) 20–29

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Characterizing heavy metal build-up on urban road surfaces: Implication for stormwater reuse An Liu a,b, Liang Liu a, Dunzhu Li a, Yuntao Guan a,c,⁎ a b c

Research Centre of Environmental Engineering and Management, Graduate School at Shenzhen, Tsinghua University, 518055 Shenzhen, People's Republic of China Cooperative Research and Education Centre for Environmental Technology, Kyoto University–Tsinghua University, 518055 Shenzhen, People's Republic of China School of Environment, Tsinghua University, Beijing 100084, People's Republic of China

H I G H L I G H T S • Heavy metal (HM) build-up varies with traffic and road surface conditions. • Traffic congestion and surface roughness exert a higher impact on HM build-up. • A “fit-for-purpose” strategy could suit urban road stormwater reuse.

a r t i c l e

i n f o

Article history: Received 7 October 2014 Received in revised form 6 February 2015 Accepted 8 February 2015 Available online xxxx Editor: Simon Pollard Keywords: Stormwater quality Stormwater pollutant processes Heavy metal build-up Stormwater reuse Ecological risk assessment

a b s t r a c t Stormwater reuse is increasingly popular in the worldwide. In terms of urban road stormwater, it commonly contains toxic pollutants such as heavy metals, which could undermine the reuse safety. The research study investigated heavy metal build-up characteristics on urban roads in a typical megacity of South China. The research outcomes show the high variability in heavy metal build-up loads among different urban road sites. The degree of traffic congestion and road surface roughness was found to exert a more significant influence on heavy metal build-up rather than traffic volume. Due to relatively higher heavy metal loads, stormwater from roads with more congested traffic conditions or rougher surfaces might be suitable for low-water-quality required activities while the stormwater from by-pass road sections could be appropriate for relatively high-water-quality required purposes since the stormwater could be relatively less polluted. Based on the research outcomes, a decisionmaking process for heavy metals based urban road stormwater reuse was proposed. The new finding highlights the importance to undertaking a “fit-for-purpose” road stormwater reuse strategy. Additionally, the research results can also contribute to enhancing stormwater reuse safety. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Stormwater is receiving significant attention as a viable alternate water resource for reuse which is currently under-utilized (Al-Salaymeh et al, 2011; Shannak et al., 2014). This is particularly important to water-scarce regions and areas where rainfall characteristics such as rainfall pattern and rainfall frequency are predicted to be influenced by climate change (Tang et al., 2013). Additionally, reusing stormwater is important in the context of creating water sensitive human settlements, where urban areas act as water supply catchments with access to water from a diversity of sources, including stormwater (Floyd et al., 2014). In this context, understanding stormwater pollutant processes to ensure safe reuse is particularly important due to the fact that there are a range ⁎ Corresponding author at: Research Centre of Environmental Engineering and Management, Graduate School at Shenzhen, Tsinghua University, 518055 Shenzhen, People's Republic of China. E-mail address: [email protected] (Y. Guan).

http://dx.doi.org/10.1016/j.scitotenv.2015.02.026 0048-9697/© 2015 Elsevier B.V. All rights reserved.

of toxic pollutants such as heavy metals present in stormwater runoff (Brown and Peake, 2006; Mahbub et al., 2010). Past researchers have identified urban roads as a primary pollutant source to stormwater runoff (Wei and Yang, 2010; Adachi and Tainosho, 2004; Johansson et al., 2009). Since road stormwater commonly contains toxic pollutants (such as traffic-related heavy metals), its reuse is potentially relevant to purposes which do not require high water quality such as street clean, public toilet flushing, landscape, roadside plant irrigation and river recharge. Additionally, these toxic pollutants could potentially pose ecological risks, which vary with pollutant types, pollutant loads, pollutant toxicity and pollutant mobility. Therefore, for some reuse purposes such as plant irrigation and river recharge, it should take the potential ecological risk into account while it could not be necessarily concerned for other purposes such as street clean and public toilet flushing. Furthermore, since urban road runoff could be highly variable with road site characteristics such as traffic and road surface conditions, how to appropriately reuse them according to their quality characteristics

A. Liu et al. / Science of the Total Environment 515-516 (2015) 20–29

and potential ecological risk becomes an essential question. This requires an in-depth understanding of pollutant availability on road surfaces, namely build-up. Pollutant build-up represents pollutant availability for wash-off by runoff. Additionally, sufficient runoff volumes, which are generated by relatively large rainfall events are necessary for stormwater reuse. In this case, most of pollutants built-up on the road surfaces can be washed-off. Therefore, it was considered that pollutant build-up is able to indicate the pollutant amount present in the stormwater runoff. In other words, pollutants that are built-up are eventually washed-off by runoff and hence have a direct influence on pollutant species and concentrations in the stormwater. This in turn has an important influence on stormwater reuse strategies. For example, which road site/s' stormwater runoff are suitable for reuse? What reuse purposes are the stormwater suitable for? In this context, only build-up characteristics were investigated in the research study rather than both build-up and wash-off characteristics. Heavy metals have been of concern in the stormwater reuse due to their potential toxicity (Yi et al., 2011). This paper presents the outcomes of an extensive research study on heavy metal build-up on a range of urban roads in a typical megacity of South China. The primary objectives were: (1) to characterize heavy metal build-up loads; (2) to analyze the key factors which influence heavy metal build-up; (3) to

21

assess potential ecological risk posed by these heavy metal loads; and (4) to develop a decision-making process for heavy metals based road stormwater reuse. The new finding is expected to contribute to applying adequate stormwater reuse strategy and hence enhancing the safety of urban road stormwater reuse.

2. Materials and methods 2.1. Study sites The research study was conducted in South China, which is a subtropical zone with abundant annual rainfall (ranging from 1400 to 2000 mm) but usually suffers from water scarcity. Shenzhen, which is a typical megacity in South China (the extent of about 2000 km2 and a population of over 11 million), was selected as the study sites. In order to investigate heavy metal build-up, ten urban road surfaces in Shenzhen encompassing different land uses, traffic characteristics and road surface conditions were selected as study sites. The ten roads are primarily paved with asphalt since asphalt paved road surfaces are the typical road type in China. There are no other pollutant sources such as industrial plants close to the selected road surfaces. This ensures

Fig. 1. Selected study road sites and their characteristics.

22

A. Liu et al. / Science of the Total Environment 515-516 (2015) 20–29

that the heavy metal loads are primarily related to traffic activities. Fig. 1 shows the selected study roads and their characteristics. 2.2. Build-up sample collection Build-up samples were collected from 1.5 m × 2 m plots using a dry and wet vacuum system to collect both, coarse and fine particulates. The boundary of the plot was demarcated by using a frame. Since the composition of build-up can vary across the road surface (Deletic and Orr, 2005), the plot was selected at the middle of each road surface to maintain the consistency of the build-up sampling. This ensures to collect the representative samples. The validity of the collection methodology has been confirmed in previous research studies (for example Mahbub et al., 2010). The antecedent dry days at sample collection time was 14 days. This ensured that the build-up load had reached equilibrium. As noted by Egodawatta (2007), the build-up load approaches a constant value with the increase in antecedent dry days and after about 7 dry days it asymptotes to an almost constant value while Wicke et al. (2012) also noted that the heavy metal build-up reaches the equilibrium condition after 6 antecedent dry days. One plot build-up sample was collected from each road surface. It is hypothesized that one plot can represent the selected road site since the plot size selected ensured that the road dust built-up was homogeneous. Accordingly, a total of 10 build-up samples were collected from the 10 selected road surfaces. Considered as primary factors influencing heavy metal build-up, parameters representing traffic characteristics and road surface condition for each study road were also collected at the road site where the buildup sample collection was undertaken. These parameters included traffic volume (TV), heavy-duty vehicle volume (HV), light-duty vehicle volume (LV), traffic congestion coefficient (TC) representing the degree of traffic congestion, the number of traffic lanes (TL) and the road surface texture depth (RTD). The higher RTD value represents the rougher road surface with more holes and depressions. The collected data is given in Table 1 along with the data collection methodology. 2.3. Laboratory testing The samples collected were separated into five particle size ranges by wet sieving prior to laboratory testing. These size ranges were: N300 μm, 300–150 μm, 150–100 μm, 100–75 μm and b75 μm. For each build-up samples, total solid load (TS), solid load on the five particle sizes and particle size distribution were tested. Date relevant to solids was provided in the Supplementary Information. Additionally, all build-up subsamples were tested for copper (Cu), zinc (Zn), lead (Pb), cadmium (Cd), chromium (Cr) and nickel (Ni) as these are the common

metal species which are traffic generated as identified in research literature (Zhao et al., 2011). The testing was undertaken by a Perkin-Elmer Optima 7300 DV ICP-OES instrument after the samples were digested using the nitric acid digestion procedure specified in Standard Methods for the Examination of Water and Wastewater (APHA, 2005). Each sample measurement was performed in triplicate and the average value was used. Certified geochemical soil reference materials (GBW07430GSS16) were used in this research for QA/QC purposes. The recoveries were 75.6–102.3%, which is within the acceptable recovery range of 75–120% (Herngren et al., 2006). 2.4. Pollution level and ecological risk assessment approaches 2.4.1. Pollution level assessment approach Pollution levels of heavy metals on urban roads were estimated using the geoaccumulation index (Igeo) introduced by Muller (1969). This index evaluates heavy metal pollution by comparing build-up loads on urban roads with the original amounts in soil. The coefficient, 1.5, was selected because of possible variations in background values for a given metal in the environment as well as very small anthropogenic influences (Zhao and Li, 2013). This geoaccumulation index was calculated according to Eq. (1). 

Ci



I geo ¼ log 2 1:5Cr

ð1Þ

where, Ci Cr

Measured total heavy metal build-up load (mg/g) Background value in soil (mg/g), 0.0271 mg/g, 0.079 mg/g, 0.0312 mg/g, 0.000225 mg/g, 0.0685 mg/g and 0.0296 mg/g for Cu, Zn, Pb, Cd, Cr and Ni respectively (Chen et al., 2015). The background value in soil used was obtained from a previous research study, where the national soil survey was undertaken across the mainland of China from April 2005 to December 2013, collecting over 38 thousand soil samples (Chen et al., 2015). The background values were considered as the amounts of heavy metals existing in the natural environment without influences exerted by anthropogenic activities. According to Muller (1969), the Igeo index value for a specific metal can be classified as: unpolluted (if Igeo ≤ 0); unpolluted to moderately polluted (if 0 b Igeo ≤ 1); moderately polluted (if 1 b Igeo ≤ 2); moderately to heavily polluted (if 2 b Igeo ≤ 3); heavily polluted (if 3 b Igeo ≤ 4); heavily to extremely polluted (if 4 b Igeo ≤ 5) and extremely polluted (if Igeo ≥ 5).

Table 1 Parameters related to road site characteristics. Road ID

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 a

Geo-coordinates N

E

22° 34′ 1″ 22° 32′ 22″ 22° 31′ 39″ 22° 30′ 55″ 22° 30′ 41″ 22° 31′ 18″ 22° 32′ 14″ 22° 33′ 37″ 22° 33′ 56″ 22° 35′ 36″

113° 55′ 44″ 113° 55′ 30″ 113° 55′ 28″ 113° 53′ 49″ 113° 55′ 27″ 113° 56′ 6″ 113° 57′ 6″ 113° 57′ 11″ 113° 57′ 11″ 113° 58′ 54″

TVa(v/h)

LVb(v/h)

HVc(v/h)

RTDd (mm)

TCe (h/km)

TLf

4900 7400 1300 2300 310 600 4000 3700 2800 500

3800 6700 1200 750 280 500 3600 3300 2400 390

1100 700 100 1550 30 100 400 400 400 110

1.324 0.998 1.506 0.371 1.102 1.499 1.916 1.330 0.990 1.468

0.030 0.030 0.040 0.022 0.038 0.030 0.026 0.022 0.026 0.026

6 8 6 8 2 4 8 6 6 4

Traffic volume per hour; collected by manual counting, equals to the summation of LV and HV. Light-duty vehicle volume per hour. c Heavy-duty vehicle volume per hour. d Road texture depth, which affects pollutant build-up on road surfaces and the interaction between the vehicle tires and the driving surface; collected using the Sound Patch Method, according to the recommendations of the US Federal Highway Administration (US Federal Highway Administration, 2005). e Traffic congestion coefficient, represented by the reciprocal of average vehicle speed (km/h). It was assumed that lower vehicle speed indicates more congested traffic. The average vehicle speeds on each road were obtained from the Transport Commission of Shenzhen Municipality. f No. of traffic lanes. b

A. Liu et al. / Science of the Total Environment 515-516 (2015) 20–29

0.8

Pj

Standard deviation for all study sites

Percentage of solids with particle size j accounting for the total TS load Particle size response factors for potential mobility in runoff. b75 μm = 13; 75–100 μm = 4.5; 100–150 μm = 4.3; 150– 300 μm = 2.2 and N 300 μm = 1 (Zhao and Li, 2013).

Mj

Heavy metal loads (mg/g)

0.6

Mean values for all study sites 0.4

3. Results and discussions 3.1. Preliminary analysis

0.2

0.0 Cu

Zn

Pb

Cd

Cr

Ni

Fig. 2. Heavy metal build-up loads.

2.4.2. Ecological risk assessment approach Ecological risk assessment of heavy metals was undertaken using the approach developed by Zhao and Li (2013), which was modified based on the study conducted by Håkanson (1980). The modified approach was for heavy metal built-up on the road surfaces, where heavy metal loads, toxic response factor and TS mobility due to the runoff wash-off are taken into account. It is noteworthy that the ecological risk assessment is primarily for water environment. Therefore, relevant parameters assigned such as toxic response factor are for water environment. For different cases such as plant irrigation, these parameters could be selected differently. This approach is represented as Eq. (2). According to Håkanson (1980), Eir low risk is defined by the highest Tir in the investigated heavy metals. In this study, the highest Tir is 30 for Cd. Therefore, potential ecological risk is classified as: low (Eir ≤ 30); moderate (30 ≤ Eir b 60); considerable (60 ≤ E ir b 120); high (120 ≤ E ir b 240) and very high (Eir ≥ 240).

i

Er ¼

 m  X Cj i Tr   A  P j  M j Cr j

ð2Þ

where, Eir Tir

Cj Cr A

23

Ecological risk factors for heavy metal i toxic-response factor, the values for each heavy metal are in the order of Zn = 1 b Cr = 2 b Cu = Ni = Pb = 5 b Cd = 30 (Håkanson, 1980) Measured loads of heavy metal in TS with particle size j Same as the definition in Eq. (1) TS loads per unit area factor. The values were divided into six levels: 0–30 g/m2 = 1.00; 31–60 g/m2 = 1.75; 61–90 g/ m2 = 2.50; 91–140 g/m2 = 3.00; 141–190 g/m2 = 3.50; N190 g/m2 = 3.75 (Zhao and Li, 2013).

3.1.1. Comparison of heavy metal build-up loads Heavy metal build-up loads on urban roads were initially compared in order to gain an initial understanding of heavy metal build-up characteristics. Fig. 2 shows the mean values and standard deviations of heavy metal build-up loads for the ten study sites. It can be noted that Zn (0.432 mg/g) has the highest mean value, around ten times higher than other heavy metals, followed by Cu (0.061 mg/g), Ni (0.045 mg/ g), Pb (0.015 mg/g) and Cr (0.005 mg/g) while Cd (0.001 mg/g) has the lowest value. Additionally, the standard deviation value for Zn build-up load was also the highest (0.217 mg/g) among the six heavy metals, which implies high variability of Zn build-up load for the different road sites. Since the build-up samples were collected from ten road surfaces with different traffic characteristics and road surface conditions, these results mean that Zn is the most dominant heavy metal on urban roads and is the metal species that is the most significantly influenced by road site specific characteristics such as traffic activities. Table 2 compares the build-up loads obtained from this research study to other cities in China. It is evident that Zn generally has the highest build-up load on urban roads compared to other metals in other cities, although the magnitudes are different, thus confirming the results of this study. Furthermore, it can be noted that heavy metal build-up loads are highly variable nationwide. High metal loads on urban roads are generally observed in large cities such as Beijing and Shanghai, while the corresponding values in Shenzhen (this research study) are relatively lower. This can be attributed to the larger traffic volumes in the more populated cities. 3.1.2. Pollution levels The Igeo index values for the six heavy metal species for the ten study roads are shown in Fig. 3. It is evident that all values of Pb (except for Road 2) and Cr are negative while the corresponding values for the other four heavy metals (Cu, Zn, Cd and Ni) at most study roads are positive. Furthermore, the Igeo index values of Zn and Cd are relatively higher, ranging from 0.192 to 2.615 and from −0.138 to 2.045, respectively, while Cu and Ni range from −0.456 to 1.235 and from −1.751 to 0.720, respectively. Therefore, it can be considered that Zn, Cd, Cu and Ni build-up on urban roads in Shenzhen are at the relatively moderately to heavily polluted levels while Pb and Cr are at the unpolluted level. These observations imply that Cu, Zn, Ni and Cd are primarily related to anthropologic activities such as traffic activities since these four heavy metals are at the moderately to heavily polluted levels while the relatively lower pollution levels of Pb and Cr mean that these could be

Table 2 Total heavy metal build-up loads in different cities, China (mg/kg). City

Population (million)

Cu

Zn

Pb

Cd

Cr

Ni

References

Shenzhen Zhenjiang Beijing Hongkong Guangzhou Xi'an Shanghai Nanjing Urumqi

11 3 21 7 13 9 25 8 4

29.6–95.7 158 552.3 173 176 20–1071 17–1175 28.7–272 33.63–252.17

135.4–725.7 686 2497.7 1450 586 80–2112 82–2136 140–798 69.26–846.15

4.8–52.8 589 595.6 181 240 29–3060 28–4443 37.3–204 13.87–99.45

0.1–1.4 N/A 2.6 3.77 2.41 N/A 0.36–4.72 0.44–2.19 0.11–14.57

1.6–12.2 129 426.8 N/A 78.8 28–853 18–1325 60.6–250 20.23–174.57

13.2–73.2 125 166 N/A 23.0 N/A 8–1251 24.8–268 19.40–95.55

This study Zhu et al. (2008) Zhao et al. (2010) Li et al. (2001) Duzgoren-Aydin et al. (2006) Han et al. (2006) Shi et al. (2008) Hu et al. (2011) Wei et al. (2009)

24

A. Liu et al. / Science of the Total Environment 515-516 (2015) 20–29

Igeo value 4 3

R1

2

R2

1

R3 R4

0

Cu

-1

Zn

Pb

Cd

Cr

R5

Ni

R6

-2

R7

-3

R8

-4

R9 R10

-5 -6 Fig. 3. Igeo index values for heavy metals.

related to roadside soil in the urban environment. As leaded fuel was phased out in Shenzhen in 1998, Pb detected could be primarily due to the past vehicle emissions (Egodawatta et al., 2013).

overshadowing critical relationships between heavy metal build-up and their influential factors (Egodawatta and Goonetilleke, 2006). PCA is an effective technique to explore correlation among variables and objects (Kokot et al., 1998). The number of significant principal components was selected using the Scree plot method (Adams, 1995). StatistiXL software (StatistiXL, 2007) was used for PCA. The detailed information regarding PCA is provided in the Supplementary Information. For PCA, objects included the five particle size fractions for the ten study roads while the variables were the six heavy metal loads in build-up for the individual size ranges, the five traffic parameters (TV, HV, LV, TL and TC) and the road surface condition parameter (RTD).

3.2. Investigation of heavy metal build-up and influential factors 3.2.1. Analysis of all particle size fractions The dataset for the five particle size fractions (N300 μm, 300– 150 μm, 150–100 μm, 100–75 μm and b 75 μm) was initially analyzed using principal component analysis (PCA). This was primarily to identify appropriate influential factors and to avoid correlating parameters

5ii 5iii5v 5iv

3

6v 6iii 66ii iv 10iii 10ii 10iv 10v 3ii 3iv 3iii 3v

2

TC

5i

RTD 10i

1

9ii 9iii 9v 9iv 8v 8ii 8iv 8iii 7v 7iv 7iii 7ii

PC 2 (27.0%)

0

4ii 4iii 4iv

-1

-2

3i

1ii 1iii 1iv 1v 4v

6i

7i 4i

Ni 8i

9i

Cu Cd Zn

1i

2ii 2iii 2iv

Pb

Cr

-3

2v -4

HV

-5

2i

LV

TL TV

-6 -3

2

7

PC 1 (39.6%) Fig. 4. PCA biplot for all particle sizes. (TV = traffic volume; HV = heavy-duty vehicle volume; LV = light-duty vehicle volume; TC = traffic congestion coefficient; RTD = road texture depth; and TL = No. of traffic lanes; v = N 300 μm; iv = 300–150 μm; iii = 150–100 μm; ii-100–75 μm and i = b75 μm; digitals represent Road ID).

A. Liu et al. / Science of the Total Environment 515-516 (2015) 20–29

Accordingly, a data matrix of 50 × 12 was submitted to PCA and the resulting biplot is shown in Fig. 4. It can be observed that all objects are separated into two groups based on particle sizes. Objects representing b75 μm are located at the positive PC1axis while nearly all of other particle size objects (N300 μm, 300–150 μm, 150–100 μm and 100–75 μm) are on the negative PC1 axis. Additionally, all heavy metal vectors as well as RTD and TC vectors are also projected on the positive PC1 axis while TV, HV, LV and TL vectors are on the negative PC1 axis. Furthermore, it is also evident that b75 μm objects are closely related to RTD and TC vectors since they are positioned at the positive PC1 axis. This could be due to the fact that vehicle emission is primarily relevant to finer particles, which are easily fixed into the rougher surfaces. These observations imply that higher heavy metals loads are attached to finer particles such as b75 μm. This is due to the fact that smaller particles have a relatively higher specific surface area (Gunawardana et al., 2014). Furthermore, it is noted that TV, HV, LV and TL vectors form an acute angle representing a close relationship between traffic volume and the number of traffic lanes. However, although RTD and TC vectors also show a small angle, these two parameters represent different characteristics of road sites (road surface condition and degree of traffic congestion, respectively). Additionally, according to the correlation matrix (see Table 3A), the correlation coefficients of TV–TL (0.727), TV–LV (0.978), TL–HV (0.639) and TL–LV (0.642) are high while the correlation coefficient of RTD-TC is relatively low (0.233). Therefore, TV, RTD and TC were included in the following analysis while TL, HV and LV were removed from further analysis.

25

3.2.2. Detailed analysis As discussed in Section 3.2.1, N75 μm and b 75 μm particle sizes show the different behaviors since in the PCA biplot (Fig. 4), objects representing N75 μm and b 75 μm are located at negative and positive PC1 axis respectively. Therefore, the further investigation into heavy metal build-up and their influential factors was conducted based on particle size fractions of N 75 μm and b75 μm, separately. This was to avoid the influence of particle sizes on heavy metal build-up. PCA was undertaken for the two datasets of N75 μm and b75 μm as shown in Fig. 5. For the PCA biplot for N75 μm (data matrix of 40 × 9), objects included the four particle size ranges (N 300 μm, 300–150 μm, 150–100 μm and 100–75 μm) for the ten study roads while the variables were the six heavy metal build-up loads for the four particle sizes, two traffic parameters (TV and TC) and the road surface condition parameter (RTD). In the b75 μm PCA biplot (data matrix of 10 × 9), objects were the b75 μm particle size for the ten study roads while the variables were the six heavy metal build-up loads for b 75 μm, two traffic parameters (TV and TC) and the road surface condition parameter (RTD). 3.2.2.1. Heavy metal build-up on N75 μm particle size. As shown in Fig. 5a, heavy metal build-up load vectors are clustered into two groups. Group 1 includes Zn, Cu, Cd and Ni vectors along with TC vector and RTD vector while Group 2 consists of Cr and Pb vectors. However, TV vector forms an obtuse angle with all of the heavy metal vectors. The two groups formed imply different influential factors for heavy metal build-up. Cu, Zn, Ni and Cd attached to N75 μm particles are primarily contributed by traffic activities while Pb and Cr would be primarily related to the

Table 3 Correlation matrices. A

Cu Zn Pb Cd Cr Ni RTD TC TV TL LV HV

Cu

Zn

Pb

Cd

Cr

Ni

RTD

TC

TV

TL

LV

HV

1.000 0.964 0.420 0.938 0.858 0.859 0.040 0.205 −0.122 −0.202 −0.083 −0.210

1.000 0.413 0.900 0.898 0.787 0.033 0.222 −0.055 −0.140 −0.016 −0.181

1.000 0.379 0.375 0.329 −0.077 0.087 0.234 0.080 0.259 −0.014

1.000 0.821 0.781 0.111 0.131 −0.058 −0.133 −0.014 −0.207

1.000 0.622 −0.049 0.125 0.186 0.086 0.206 −0.012

1.000 0.191 0.070 −0.173 −0.152 −0.131 −0.239

1.000 0.233 −0.084 −0.171 0.066 −0.663

1.000 −0.271 −0.461 −0.176 −0.505

1.000 0.727 0.978 0.481

1.000 0.642 0.639

1.000 0.287

1.000

B

Cu Zn Pb Cd Cr Ni RTD TC TV

Cu

Zn

Pb

Cd

Cr

Ni

RTD

TC

TV

1.000 0.877 0.162 0.526 0.069 0.767 −0.082 0.450 −0.507

1.000 0.246 0.591 0.353 0.750 −0.014 0.644 −0.401

1.000 0.034 0.401 −0.047 −0.125 0.064 0.284

1.000 0.111 0.564 0.248 0.351 −0.498

1.000 −0.104 −0.127 0.427 0.493

1.000 0.182 0.374 −0.597

1.000 0.233 −0.084

1.000 −0.271

1.000

Cu

Zn

Pb

Cd

Cr

Ni

RTD

TC

TV

1.000 0.910 0.955 0.928 0.661 0.470 0.277 0.490 −0.033

1.000 0.903 0.762 0.717 0.222 0.145 0.649 −0.048

1.000 −0.149 −0.153 0.363 0.571

1.000 0.729 0.074 −0.388

1.000 0.233 −0.084

1.000 −0.271

1.000

C

Cu Zn Pb Cd Cr Ni RTD TC TV

1.000 0.889 0.633 0.431 0.277 0.389 0.016

1.000 0.587 0.390 0.241 0.250 0.104

26

A. Liu et al. / Science of the Total Environment 515-516 (2015) 20–29

traffic or rougher surfaces needs extra care due to the relatively higher heavy metal levels.

urban roadside soil. This is in agreement with the conclusions derived in Section 3.1.2. Furthermore, TC, RTD and the heavy metal vectors forming one group imply that compared to traffic volume, the degree of traffic congestion and road surface roughness could have a more important influence on heavy metal build-up. This can be also supported by the correlation matrix (Table 3B). The correlation coefficient values of TCheavy metals and RTD-heavy metals (except for Cr) are higher than the corresponding values of TV-heavy metals. The slower moving traffic as a result of congestion contributes significantly to heavy metal buildup although the traffic volumes may not significantly change due to incomplete fuel combustion and increased exhaust emissions as well as more frequent go-and-stop activities resulting in greater tire and brake lining wear (Thorpe and Harrison, 2008; Mahbub et al., 2010). Higher road texture depth leading to higher heavy metal build-up can be attributed to rougher road surfaces resulting in increased tire wear and hence higher heavy metal build-up loads. Additionally, rougher road surfaces allow more particulates to be retained on the surface despite the occurrence of particulate removal processes such as rainfall events and wind turbulence. These conclusions suggest that reusing stormwater from road sites with more congested 7

a: >75 µm (>300 µm, 300-150 µm, 150-100 µm and 100-75 µm)

6

3.2.2.2. Heavy metal build-up on b 75 μm particle size. As shown in Fig. 5b, the clustering of heavy metal vectors is different compared to the N75 μm PCA biplot. All heavy metal build-up vectors are projected on the positive PC1 axis along with RTD and TC vectors while TV vector points to the direction of the negative PC2 axis. These results confirm that traffic volume is secondary to the degree of traffic congestion and road surface roughness in influencing heavy metal build-up even for the finer particle size fraction. It is also noteworthy that Ni build-up vector forms a very small angle with RTD (this can be also supported by the correlation matrix in Table 3C, where the correlation coefficient of Ni-RTD is 0.729) and an acute angle with TC while Cu, Pb, Zn and Cd vectors are close to each other and also have acute angles with TC. However, Cr vector points away from the other heavy metals as identified in Fig. 5a. These observations suggest that the heavy metal build-up characteristics for the finer particle sizes (b 75 μm) are different from the case of N 75 μm, particularly for Pb. Different from the case of N75 μm, Pb build-up also shows a relationship with traffic activities such as traffic congestion in the case of b 75 μm. This implies that traffic activity also exerts influence

2v

Group 2

5

Cr

PC 2 (21.9%)

4

2ii2iii 2iv

TV

3

Pb

Group 1

2 1

3iii 3iv 3ii

1iii 1iv 1v 1ii 9iii 9v 9ii 8iv 8v 8ii 4iii4iv 4ii 8iii 7v 7iv 7iii 7ii

0 -1

TC

3v

Zn

5iv 5ii 5iii 4v Cu5v 6v 6iii 6iv 9iv 6ii 10vCd 10iii 10ii10iv

Ni

RTD

-2 -4

-2

0

2

4

6

PC 1 (42.4%)

2

b:

Characterizing heavy metal build-up on urban road surfaces: implication for stormwater reuse.

Stormwater reuse is increasingly popular in the worldwide. In terms of urban road stormwater, it commonly contains toxic pollutants such as heavy meta...
1MB Sizes 2 Downloads 6 Views